TechTurco WhatsApp ChannelOne-way channel: your number stays private
Artificial Intelligence
Google announces Gemini Omni 1.2 Flash: Price drops 30%, video length increases to 15 seconds
Developed by Google DeepMind, Gemini Omni 1.2 Flash has been released with 15-second video generation capabilities and a 30% reduced usage cost. Here are the details of the features it offers.
Contents
Google made its Gemini Omni 1.2 Flash model generally available on the Gemini API and enterprise platforms. According to information shared on November 5, 2026, the new artificial intelligence (AI) model extends video generation lengths while lowering processing costs.
Pricing and Workflow Costs
The new model features a 30% discount compared to the Omni 1.1 version, pricing 720p resolution video at $0.07 per second. The system provides a cost-efficient workflow. Users can initially create quick drafts at 360p or 720p resolution. Specific details can then be edited via chat prompts without regenerating the full video. Selected videos can finally be upscaled to 1080p or 4K resolution.
Video Length Increased to 15 Seconds and Reference Fidelity
The direct video generation limit, previously 10 seconds in the earlier version, was increased to 15 seconds. Users can set the generated video length between 3 and 15 seconds in one-second increments. The model can reference character, product, or style images, as well as short motion clips up to three seconds and audio cues.
Reklam
Facial features, clothing, typography on packaging, and brand colors are preserved across different camera angles. In an example use case, combining a model photo with front and back product images of three different outfits allowed Gemini Omni 1.2 Flash to generate 15-second try-on videos with a 360-degree rotation while maintaining the model's identity and clothing cut details.
Chat-Based Editing and Audio Synchronization
The multi-turn conversational editing feature allows step-by-step modifications on the same video. Changing the background, adjusting lighting or weather conditions, and altering the aspect ratio are possible without regenerating the entire clip. For example, in a generated clip, the background can be sequentially changed to a tent, oak forest, horse barn, fairground, and stone castle while preserving character movements, armor reflections, and camera angles.
The model also provides more natural dialogue, clear lip sync, and multi-speaker separation in character speeches. These capabilities can be used in avatar and voiceover processes across Flow, Gemini for business, and Google Vids.
Reklam
Be the first to know
Join our channel and get key developments straight on your phone.
TechTurco Telegram ChannelMute or leave any time
Camera Controls and Platform Access
The new version reduced unwanted cuts in complex camera movements such as FPV drone flights, zooming, and panning. Guided by two architectural reference photos, the model can generate a continuous drone shot starting from a marketplace, passing through bronze doors, and reaching the Pantheon dome.
Google stated that development work continues on fast sports movements, sudden scenario changes, and complex physics simulations. An imperceptible SynthID watermark is included in all generated videos. The model is available on the following platforms:
- Gemini API and Google AI Studio for developers
- Google Cloud console for enterprise users
- Flow, Google Vids, Google Ads, and the Gemini app for Google AI Ultra subscribers for content creators and subscribers
Reklam
This article was translated from Turkish. Read the original article in Turkish.
More in Artificial Intelligence
